Intended for healthcare professionals


Hormone replacement therapy and tumour grade in breast cancer: prospective study in screening unit

BMJ 1996; 312 doi: (Published 29 June 1996) Cite this as: BMJ 1996;312:1646
  1. Claudia Harding, research registrar in surgerya,
  2. W F Knox, consultant histopathologista,
  3. E B Faragher, medical statisticiana,
  4. A Baildam, senior lecturer in surgerya,
  5. N J Bundred, senior lecturer in surgerya
  1. a University Departments of Surgery and Histopathology, University Hospital of South Manchester, West Didsbury, Manchester M20 2LR
  1. Correspondence to: Mr Bundred.
  • Accepted 29 February 1996

Several studies report an increased incidence of breast cancer in women taking hormone replacement therapy (apparently a duration dependent phenomenon)1 2 3; however, a decreased mortality from breast cancer has been reported in women using the treatment.1 We assessed the prognostic factors for breast cancers detected by screening at one unit to determine if cancers developing in women who used hormone replacement therapy have a better prognosis.

Patients, methods, and results

All postmenopausal women (>12 months since last menstrual period or women over 50 who had had a hysterectomy) with invasive breast cancer detected at our screening unit between 1989 and 1994 were identified. Prognostic factors and history of use of hormone replacement therapy were recorded. Women taking hormone replacement therapy for more than three months within the year before their breast cancer diagnosis were considered “users.” After definitive surgery, pathological tumour size, grade, axillary node metastases, and steroid receptor status were compared between users of hormone replacement therapy and non-users.

Of the 433 postmenopausal women with screen detected cancer, fewer women diagnosed in the prevalence round than in the incident (second) screen were using hormone replacement therapy (74/325 (23%) v 34/108 (32%); χ2 = 3.3; P>0.05). Median duration of use was 24 (range 3-264) months.

Well differentiated grade I cancers occurred more often in users of hormone replacement therapy than non-users (45% v 20%; P<0.001; table 1); 15% (16) of users compared with 45% (146) of non-users had grade I, well differentiated tumours (P<0.001). The likelihood of cancers detected in the users being well differentiated increased from the prevalence screen (relative risk 2.19; 95% confidence interval 1.54 to 3.12) to the incident screen (2.49; 1.38 to 4.49). Tumour size and steroid receptor status (analysed by immunocytochemical assay) were similar in the groups. The excess of grade I tumours was seen equally in the 72 women using oestrogen only and the 36 using a combined preparation.

Table 1

Tumour grade of breast cancers in users and non-users of hormone replacement therapy

View this table:

In women undergoing axillary surgery, node positivity occurred in 23 (26%) of users and 68 (28%) of non-users (difference -2.6%; -14% to 8.5%). In the total group, positivity was 21% in users and non-users (69/325 v 23/108).

A total of 28 women developed recurrent cancer after a median follow up of 45 (range 12-96) months. Recurrence was similar in users and non-users (8 (7%) v 20 (6%); P = 0.99, log rank test).


Although an increased breast cancer risk (up to twofold) has been linked with long term use of hormone replacement therapy, a large British cohort study has shown that mortality from breast cancer in hormone replacement therapy users is almost halved,1 a finding supported by a subsequent American case-control study.2 The American study included both screen detected and symptomatic cancers, with significantly more screen detected cancers included among cases than controls, which would account for the differences in node positivity and mortality between the two groups. By studying only screen detected cancers we eliminated such surveillance or ascertainment bias.

Tumour grade is important for predicting survival from breast cancer, and prognostic indices usually include grade.4 Among screen detected cancers, hormone replacement therapy users had better differentiated tumours than non-users, which may explain their reduced mortality.1 5 In both screening rounds, women who used hormone replacement therapy had better grade tumours. As users and non-users were screened equally in the screening rounds, this effect is not due to ascertainment bias. Combined hormone replacement therapy (containing progestogens) has recently been claimed to increase the risk of breast cancer,3 but we found that better grade tumours developed with both types of treatment. Although our follow up remains short, no difference in recurrence has been seen between the two groups.

Prescribing hormone replacement therapy after breast cancer treatment remains controversial, but its use at the time of breast cancer diagnosis is unlikely to effect mortality adversely. Therefore hormone replacement therapy after breast cancer treatment should not be denied to patients if warranted by symptoms and if alternatives for the control of severe menopausal symptoms have failed.

We are grateful to Dr Elizabeth Anderson for performing steroid receptor analysis and the consultant radiologists at the Nightingale Centre, Manchester, for their help in diagnosis of the patients.


  • Funding None.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract